Humans are only incidental hosts of zoonotic pathogens, such as the agent of Lyme disease (LD), which are maintained in nature in one or more reservoir host species. Therefore, rather than focusing upon humans for disease prevention, a more effective way to reduce LD risk is to decrease the number of infected ticks by reducing the prevalence of the pathogen in the wildlife reservoir hosts. Control measures targeted to reservoir hosts, such as vaccinating wildlife or applying acaricides are receiving increasing attention as control methods for LD. However, it is essential to identify the most important host species to which these methods should be targeted, since hosts vary greatly in their ability to acquire and transmit Borrelia burgdorferi, the agent of LD. There is increasing evidence that different host species transmit different B. burgdorferi genetic variants, some of which have higher pathogenicity in humans. Knowledge of which animal species are the most important in maintaining this pathogen is very limited, due to the logistical difficulty of studying mammals and birds in natural settings. This project will examine the variability in the composition of the B. burgdorferi reservoir host community across a large geographic area. Also, by coupling this study with another NIH-funded project, it will provide information on the effect of host community composition on B. burgdorferi prevalence and genetic makeup in host-seeking ticks, which are responsible for transmission of B. burgdorferi to humans. The results of this project will help design regional intervention methods for LD and possibly other vector-borne zoonoses. Two methodological approaches will be explored to facilitate host-targeted intervention strategies on such a large geographic scale. First, cost- effective indirect measures of mammalian host abundance will be used and second, predictive models of host abundance will be developed based on landscape patterns, yielding estimates of host abundance in unsampled areas. Sampling of mammals and birds will be conducted in five sites along a 400 km N-S transect in the northeastern US. Within each site, sampling will be conducted along two site transects. Along each site transect, rodents will be surveyed using tracking tubes in two grids, medium-sized mammals will be surveyed in ten scent stations and birds species will be recorded in point counts around the scent station sites. The hierarchical sampling design will allow determination of the scale of variability, that is, does the host community composition vary by site, within sites among transects or within transects? Landscape variables will be measured at the same scales in order to develop predictive models of host species composition. Finally, host community composition will be compared to the abundance of I. scapularis and the prevalence of B. burgdorferi genotypes in nymphal ticks to develop a predictive model of pathogen genotype distribution and prevalence. The results of this project will provide crucial information that will improve the effectiveness of Lyme Disease prevention methods directed towards wildlife reservoirs, such as vaccinating wildlife. This project provides a method to identify the wildlife host species that are the most important reservoirs and should therefore be targeted in efforts to reduce Lyme Disease risk to humans. Novel methods will be used to estimate wildlife species composition and ecological models will be developed to predict LD risk over a large geographic area based upon wildlife species composition.